Cold Exposure Changes the Blood Metabolism Parameters in Piglet
Cold exposure changes the body's metabolic state to maintain energy homeostasis. To explore the effects of cold exposure on the metabolism of piglets, we tested the energy metabolism parameters of blood serum after 48h of cold exposure with or without addition of broad-range antibiotics (Anti) to drinking water to deplete the intestinal microbiota. Glucose and insulin concentrations were not found to be different between each treatment (Fig. 1a and 1b). Cortisol was lower under cold exposure treatment than at room temperature (RT) (Fig. 1c, P < 0.01). Leptin and adiponectin were lower under cold exposure than under RT treatment regardless of microbiota depletion (Fig. 1d and 1e, P < 0.05). However, under the same treatment, depletion of the gut microbiota did not influence the leptin and adiponectin concentrations in blood serum (P > 0.05). The T3 concentration was inhibited by cold exposure compared with that at RT (Fig. 1f, P < 0.05). After depleting piglet intestinal microbiota with Anti, the T3 concentration was significantly lower than that under RT treatment alone, which means that the intestinal microbiota played an important role in maintaining T3 homeostasis (Fig. 1f, P < 0.05). In response to cold exposure, T4 had the same result as T3 without microbiota depletion; however, there was no difference under RT treatment with or without Anti addition (Fig. 1g). There was a tendency toward decreased GLP-1 concentrations in the blood serum of cold-exposed or RT + Anti treated piglets that bordered significance (Fig. 1h). The PYY concentration had the same trend as the GLP-1 response to cold exposure; however, there was no significant difference with or without the microbiota depletion using Anti (Fig. 1i). Together, these data suggested that cold exposure would selectively affected blood energy metabolism parameters and that most of them were inhibited. The intestinal microbiota played an important role in the same parameters under the RT environment.
Cold Exposure Increased the IgA Levels and Insulin Sensitivity
To investigate whether the cold exposure influences the weight of important organs, after slaughter, we measured the liver, spleen, and thymus weights (Fig. 2a-c). Only liver weight had a decreasing trend under cold stress compared with that at RT (Fig. 2a, P = 0.059). IgA and IL-18 in the blood serum of piglets were increased under cold exposure, while there were no significant differences between other treatments, including with or without depletion of microbiota (Fig. 2d and 2h). IgG, IL-6, IL-10, and NF-kB were not significantly different under the four treatments (Fig. 2e, 2f, 2g, and 2i). We also detected glucose tolerance by measuring the blood sugar concentration at different times after oral administration of glucose (Fig. 2j). The results showed that cold-exposed piglets without microbiota depletion had an elevated glucose peak following glucose gavage compared to that under other treatments after 40 min (Fig. 2j), but also the fastest clearance. Interestingly, there was nearly the same initial glucose peak between the cold and cold + Anti treatments. This suggested that the oral glucose was rapidly taken up in cold-exposed piglets without microbiota depletion. However, piglets subjected to intestinal microbiota depletion under room or cold temperature showed a delayed peak appearance at 60 min, and the speed of clearance was slower than that under cold treatment but higher than that under normal RT treatment. This suggested that a broad range of antibiotics that interrupted the piglet intestinal microbiota would influence glucose uptake and metabolism in the host. In our models, the duodenum mucosal thickness, villus length, crypt depth, and cecal mucosal thickness increased in the cold-exposed piglets compared with those in the RT-treated piglets (Figure S1a, 1b, 1d, and 1e). Interestingly, there was an antibiotic-dependent effect on the above parameters, which were increased upon intestinal microbiota depletion at room temperature. However, under cold exposure with antibiotic treatment, the duodenal mucosal thickness and villus length decreased compared to those under cold exposure without antibiotic treatments (Figure S1a and S1b) but duodenal crypt depth increased. These results suggest that the intestinal morphological response to the environment was also determined by intestinal microbiota homeostasis.
Cold Exposure Changes the Microbiota Composition and the Content and Epithelial Surface of the Caecum
To investigate whether the cold stress causes changes in the intestinal microbiota, we collected the cecal content and epithelium-attached samples of cold-exposed and RT-treated piglets (Fig. 3 and Figure S2). Determination of the composition by 16S rRNA marker gene sequencing, followed by principal coordinates analysis (PCoA) based on the weighted UniFrac distance, showed the large alterations in the microbiota compositions under cold exposure (Fig. 3a). We observed that the abundant phyla were Firmicutes, Bacteroidetes, Proteobacteria, Epsilonbacteraeota, and Actinobacteria. Firmicutes was the most abundant phylum in all samples (on average 72.16%), followed by Bacteroidetes (on average 23.13%) in all samples (Fig. 3b). At the family level, Lachnospiraceae, Ruminococcaceae, and Prevotellaceae were the most abundant families in all samples, accounting for 36.95%, 22.77%, and 20.43%, respectively, on average. The alpha diversity was increased after cold exposure, but only the Shannon index was significantly higher than that under the RT treatment (Fig. 3c). However, there were no significant differences found between the cold and RT treatments at the phylum level (Fig. 3d-3j). We further investigated whether the differences existed in ASVs using STAMP software and FDR adjusted P-values. We found that 39 ASVs were significantly different between the cold and RT treatments, while most of them, with 20 ASVs were higher under cold exposure than under RT treatment (Fig. 3k, P < 0.05). A total of 20 ASVs were distributed into Lachnospiraceae, Ruminococcaceae, Prevotellaceae, Rikenellaceae, Muribaculaceae. Unique ASVs were higher in the RT treatment and belonged to Eggerthellaceae, Campylobacteraceae, and Helicobacteraceae three families (Fig. 3k, P < 0.05).
The intestinal epithelium surface, where the key signal transporters, such as the short fatty acid receptor, were located, was the most important place for the interaction between intestinal microbiota and the host. Thus, we also detected the cecal epithelium-attached microbiota using 16S rRNA sequencing (Figure S2). The PCoA showed that the gut microbiota under cold exposure and RT treatment had a little overlap between samples (Figure S2a). All alpha diversities also had an increasing trend for the cold-exposed piglets, but no significant index was found (Figure S2c). The most abundant phylum in the intestinal epithelium-attached microbiota was the same as that in the cecum. However, the order of the relative abundance of these phyla was different from those in the cecum and included Bacteroidetes, Firmicutes, Epsilonbacteraeota, Proteobacteria, and Spirochaetes with averages of 50.02%, 22.44%, 17.29%, 8.73%, and 1.00%, respectively. The most abundant families were Prevotellaceae, Lachnospiraceae, Ruminococcaceae, and Muribaculaceae, at 44.29%, 9.81%, 6.55%, and 2.18%, respectively. The phyla Firmicutes and Spirochaetes showed an increasing trend compared with their levels under RT treatment (Figure S2f and S2i, P = 0.071, and 0.056, respectively). The other phyla were not influenced under cold exposure (Figure S2d, g, h, and j). At the ASV level, only 14 ASVs were found to be significantly different, 12 of which were higher than those under RT treatment (Figure S2e, P < 0.05). A total of 12 ASVs belonged to Prevotella, Alloprevotella, Ruminococcaceae, Muribaculaceae, and Erysipelotrichaceae (Figure S2e).
Cold Exposure Changes the Functions of Microbiota in the Cecum
The functions of the gut microbiome were confirmed using metagenomics sequencing. A total of 99 Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology (KO) genes in the metabolism catalog showed significant differences between cold and room conditions (P < 0.05), while 72 of these genes were more highly expressed in the cecal microbiota under cold exposure than at room temperature (Fig. 4a). Changes in gene abundances are shown with pathway representation (Fig. 4b). Among the most differentially abundant pathways, purine metabolism, arginine biosynthesis, fructose and mannose metabolism, sulfur metabolism, biotin metabolism, N-glycan biosynthesis, arginine and proline metabolism, glyoxylate and dicarboxylate metabolism, methane metabolism, oxidative phosphorylation, pentose and glucuronate interconversions, nitrogen metabolism, and glycosaminoglycan degradation were found to be associated with cold exposure (Fig. 4b). More production of purines, including adenine, hypoxanthine, xanthine, and guanine, might be associated with higher energy requirements for microbiome growth and metabolism. Gas production in the intestine is important for energy balance and even for signal transport. Under cold exposure, hydrogen, and trimethylamine for methane metabolism, ammonia for nitrogen and glyoxylate and dicarboxylate metabolism and sulfide and succinate for sulfur metabolism were higher under cold stress, which means that the microbiome might exhaust more energy to produce these products. However, porphyrin and chlorophyll metabolism were found to be significantly associated with RT conditions, and is responsible for producing vitamin B12 coenzymes (Fig. 4b). In the epithelial surface of the cecum, the total number of significantly different KO genes was only 24 under the metabolism category. Eleven of these genes were higher under cold exposure than at room temperature and were mainly (7/11) associated with amino acid metabolism (Fig. 4a). Nine of 13 KO genes were dominant at RT and related to carbohydrate metabolism and glycan biosynthesis and metabolism.
A total of 140 other KO genes were found to be significantly different between the two treatments in the cecum (P < 0.05) (Figure S3). In contrast, 56/140 KO genes were higher at room temperature, and 84/140 KO genes were dominant under cold stress (Figure S3a). For room temperature treatment, the adherence, spore germination, and motility functions at protein families: signaling and cellular processes pathways were stronger than under cold stress, which means the microbiome had more power capacity to establish or move to the proper microenvironment. For protein families: genetic information processing, 16/35 were higher under RT, and 19/35 were higher under cold stress. The functions for the cecal epithelium are shown in Figure S3b. These differences might not be just from the microbiome, but also from including the host.
Cold Exposure Elevated the Amino Acid and Bile Acid Metabolism level
To determine which metabolites are related to cold exposure in the host, we detected the blood metabolites using metabolomics. There were 133 significant metabolites in the positive model between the two treatments (Fig. 5a), while 120/133 were dominant under cold stress. The OPLS score also showed that the two treatments were significantly separated (Fig. 5b). We also used the microbe-metabolite vector (mmvec) to confirm the correlation between microbiota and metabolites according to a previous report . After analysis, we found that stearidonic acid, Arg-Thr, oxaprozin, glucosamine, His-Cys, Pro-Tyr, N-Oleoylethanolamine, D-4-Hydroxyphenylglycine, and glutamic acid were higher under cold stress than RT, and a higher correlation was observed between metabolites and the microbiome in the cecum (Fig. 8a and c). Stearidonic acid was positively correlated with the Prevotellaceae NK3B31 group, Ruminantium group, Alloprevotella, and Prevotella 9 (Fig. 8c). Oxaprozin was most correlated with Prevotella 9 and Alloprevotella. Glucosamine and His-Cys were correlated with Prevotella 9 and Ruminococcaceae UCG-005. N − Oleoylethanolamine was correlated with the Christensenellaceae R-7 group and Alloprevotella. D-4-Hydroxyphenylglycine was correlated with Prevotella 2. Phenol, His-Ala, sphinganine, and sulfanilamide were more abundant among the blood metabolites and had a high correlation with the microbiome of the mucosal surfaces of the cecum (Fig. 5a and d). Phenol was correlated with Alloprevotella and Prevotellaceae UCG-001. His-Ala was correlated with Muribaculaceae. Sphinganine and sulfanilamide were both correlated with the Prevotellaceae and Rikenellaceae RC9 gut groups.
The significant metabolites numbered 67 in the negative model between the two treatments (Figure S4a), while 28/67 were higher under the cold stress treatment. However, only a few metabolites, including 2-Dehydro-3-deoxy-D-gluconate, Anthranilic acid (Vitamin L1), and 3-Hexanone, were found to be higher under cold stress and to have a strong correlation with the microbiome. 2-Dehydro-3-deoxy-D-gluconate was strongly correlated with Prevotella 9 in cecum (Figure S4c). Anthranilic acid (vitamin L1) was correlated with Prevotellaceae UCG-003 in the mucosal surfaces of the cecum (Figure S4d). 3-Hexanone was correlated with Parabacteroides and Prevotellaceae UCG-003 in the mucosal surfaces of the cecum.
Cold stress increased the expression of UCP3 in fat
To uncover the mechanism of the microbiota-liver and fat cross-talk responsible for the energy balance, we deep sequenced the transcriptome from the liver and fat. Pathway enrichment analysis revealed that changes in pathways involved in butanoate, propanoate, and pyruvate metabolism, pentose and glucuronate interconversions, and glycolysis/gluconeogenesis related to carbohydrate metabolism were decreased in the liver under cold stress when compared to RT (Figure S5a). Glycerolipids, fatty acid degradation and metabolism, ascorbate and alternate metabolism, and retinol metabolism related to vitamins were also decreased in the liver under cold stress (Figure S5a). Tryptophan metabolism, valine, leucine, and isoleucine degradation related to amino acid metabolism and steroid hormone biosynthesis related to hormone biosynthesis were also decreased under cold stress when compared to RT (Figure S5a). Under cold stress, most signaling pathways, including cytosolic DNA-sensing, TNF, NOD − like receptor, and IL-17 signaling pathways, increased compared to those under RT. Pyrimidine metabolism, RNA degradation, RNA polymerase, proteasome, and ribosome biogenesis in eukaryotes were also increased under cold stress, compared to RT pigs (Figure S5b). The immune signaling pathways in the fat were decreased under cold stress, compared to RT pigs (Figure S5c). However, in the fat, the pathways related to the energy supply, including PPAR, cGMP-PKG, AMPK, adipocytokine, calcium signaling pathways, increased compared to those under RT. Regarding substrate metabolism aspects, aspects of energy supply, including protein digestion and absorption, vitamin digestion and absorption, glycerolipid metabolism, fat digestion, and absorption, were also increased under cold stress, compared to RT (Figure S5d). We found that the CYP8B1 involved in the main bile acid synthesis pathway was 16 times higher under cold stress in the liver than under RT (Fig. 6a). However, after treatment with mixed antibiotics, the increasing trend was suppressed, and significant differences were found between cold and cold + Anti treatments, indicating that the gut microbiota might play a very important role in manipulating the bile acid metabolism pathway (P < 0.05). Nr0b2 and FXR receptors, which could combine with bile acid to regulate the bile acid cycle, increased under cold stress in the liver compared to those under RT. The genes Adcy9, Nceh1, and Abce4, which are involved in the bile secretion pathway, also increased under cold stress. Adcy9 encodes neutral cholesterol ester hydrolase 1 responsible for transferring cholesterol ester to cholesterol, which is an important precursor for bile acid synthesis. However, we examined the genes responsible for the secondary bile acid biosynthesis in the gut microbiome, including cbd, baiB, and baiN, and they were not influenced by cold stress (P > 0.05, Figure S6). Other important microbial metabolites include short chain fatty acids, which play a very important role in host energy balance regulation through the receptors Ffar2, Ffar3, and Ffar4. We found that the genes Ffar2 and Ffar3 were upregulated by cold stress in the livers of pigs (P < 0.05, (Fig. 6a), while the Ffar4 was downregulated slightly.
The expression of genes Bcl2l1 and Mcl1 involved in antiapoptotic function increased under cold stress in the liver, compared to RT in pigs (Fig. 6b). The pro-apoptotic genes Casp6, Casp8, Casp10, and Fas also increased under cold stress. However, TNF signaling promoted apoptosis, which was activated by bacteria (Fig. 6b). Thus, these proapoptotic genes were significantly suppressed in all microbiota-depleted pigs (Fig. 6b). The FAS, Slc2a8b, and Slc2a12 genes involved in glucose transporters were upregulated under cold stress, indicating that glucose metabolism can be enhanced by the microbiota under cold stress.
Additionally, under cold stress, the insulin resistance was higher than that under RT, which was consistent with a previous insulin resistance experiment. Fat is important for thermogenesis, especially UCP genes. However, the UCP1 gene is enriched in brown adipose tissue and is absent in pigs. We also investigated whether the UCP gene involved in thermal energy was significantly different under cold stress in the liver and fat (Fig. 6c). The results revealed that the UCP2 gene counts were decreased under cold stress in the liver; in contrast, UCP3 gene counts were increased in the fat under cold stress compared to those under RT, which suggested that the fat might provide an important energy supply through UCP3 pathways for piglets under short-term cold stress. The genes Arid1b, Smarcd3, Cpt1A, Cpt1B, and Slc25a20, which are involved in fat metabolism for thermogenesis, increased under cold stress in fat (Fig. 6c). In particular, Cpt1B, which is necessary to use fat to supply energy through β-oxidation, increased under cold stress to levels 4 times those under RT. The expression of the Clps and Pnliprt1genes responsible for fat metabolism increased 18-fold under cold stress; however, no difference was found between cells treated with or without using antibiotics under cold stress (Fig. 6c).
Cold stress changed the turnover of the total, abundant, and rare ASV fractions
We calculated βNTI and RCBray to determine the assembly processes driving microbial community composition under different treatments (Fig. 7). In the total, abundant, and rare ASVs fractions, βNTI values were all between − 2 and 2, indicating that the stochastic process was critical. With further analysis, the RCBray values of the microbial communities in the abundant and most total ASVs, except RT_CeM, were >0.95, suggesting that dispersal limitation predominantly governed the microbial communities. Conversely, the RCBray values of the microbial communities of the rare ASVs were only found in C_CeM and C_CeM-RT_CeC to be > 0.95, and in the other pairwise samples, the RCBray values were < 0.95, reflecting that drifting alone (weak selection) determined the microbial community.